Artificial Intelligence between “Openness” and Open Source

The advent of Artificial Intelligence, particularly generative AI, has radically transformed the digital landscape, challenging the established principles of open and closed source. This transformation has introduced new challenges related to transparency, reusability, and accountability, necessitating updated definitions, procedures, and new types of licenses. In the European context, legislative proposals such as the AI Act aim to promote open standards but often fall short of addressing the complexity of modern AI systems. Companies, particularly in the telecommunications sector, face growing uncertainties in using and developing AI-based models, with significant implications for innovation, auditing, and transparency. This study aims to examine the impact of AI advancements with a focus on telecommunications, starting with the open/closed source dichotomy and analyzing European legislative proposals on the subject.

In recent decades, open-source software has been a cornerstone of innovation, fostering accessibility, sharing, and transparency. However, in the context of AI systems, the concept of “open source” takes on different meanings compared to traditional software, due to the proprietary nature of many algorithms and the transparency challenges related to training data, code, and model weights. European regulations, while aiming to promote open standards, leave gaps and ambiguities that can hinder the reusability and auditing of AI systems. This creates challenges for companies, especially in the telecommunications sector, which rely on AI to design advanced services. Understanding the various interpretations of open source in AI systems and aligning with regulations is therefore crucial to developing innovative and compliant solutions.

1. Examine critical legal issues related to the use of Generative Artificial Intelligence, including misuse, malfunctioning and intellectual property violations.
2. Evaluate European legislative proposals and legal activity in the US or other countries if deemed significant (e.g. China or the UK) to determine the regulatory future of Generative Artificial Intelligence.

An analysis of the main challenges of the open source paradigm in relation to the changed context of A.I, with solutions and experimental practices to overcome them.
A summary of the most recent European legislative proposals related to the development and promotion of open standards for software and for A.I. models in particular, to assess their real effectiveness and shortcomings. The analysis includes a specific section dedicated to the implications for the telecommunications services sector.